Hybrid Rule-Based, Fuzzy, and Optimization Matching Algorithm for Reserve Force Task Recommendation
ID：21
View protection：Participant Only
Updated time：2026-10-10 06:31:12
Views：22
Online
Abstract
When a reserve unit is mobilized, choosing reservists for a task takes more than finding someone with the right skill. The person also has to be active and available, have no conflicting schedule, be within a workable distance, hold a suitable rank, and not already be carrying too many assignments. In this paper we describe and test the matching algorithm used in a web-based skill-management and task-recommender system for reserve force mobilization. The algorithm works in three stages. First, a deterministic set of rules drops any candidate who cannot be assigned. Next, a fuzzy inference step turns skill match, distance, and rank difference into a single suitability score. Finally, a weighted optimization step combines suitability with rank and current workload and sorts the remaining candidates into a shortlist. We tested the algorithm with 13 mock reservist profiles across two deployment scenarios, along with 28 boundary-value unit tests, five weight settings, and three versions of the algorithm. Every unit test passed. With the full hybrid version, Precision@3 and top-three accuracy were both 100%, Recall@3 was 80%, and ranking took 0.0179 ms on average over 2,000 runs. These results suggest the method is fast and easy to interpret as a decision aid. Still, the scenarios were simulated, and the reference labels reflected criteria similar to those the algorithm uses. What the results show is internal consistency, not proof that the method works in the field.
Keywords
personnel-task matching, rule-based system, fuzzy inference, optimization, recommender system, reserve force mobilization.
Post comments